Apa yang dipikirkan agen AI tentang berita ini
The panelists debate Microsoft's valuation and growth prospects, with most acknowledging Copilot's slow adoption but arguing that Azure's growth, pricing power, and long-term backlog outweigh this concern. Key risks include the quality of Azure's remaining performance obligations, potential antitrust scrutiny, and concentration of top customers.
Risiko: The quality and composition of Azure's remaining performance obligations, particularly the extent to which it represents high-margin Microsoft software versus pass-through Azure compute costs for OpenAI/Anthropic.
Peluang: Azure's strong growth and the long-term backlog, which could provide a multi-year tailwind for Microsoft's revenue.
Microsoft baru saja menutup kuartal terburuknya di Wall Street sejak krisis keuangan 2008, karena investor merugi pada prospek raksasa perangkat lunak tersebut di bidang kecerdasan buatan.
Saham anjlok 23% pada kuartal pertama, penurunan yang lebih tajam daripada rekan-rekan technya atau Nasdaq yang turun 7% pada periode tersebut. Microsoft bangkit sedikit pada Selasa, bersama dengan pasar yang lebih luas, dengan saham naik 3,3%, lompatan terbesar sejak Juli.
Meski Microsoft tetap mendominasi perangkat lunak produktivitas tempat kerja dan melalui sistem operasi Windowsnya, perusahaan menghadapi tekanan ganda untuk tumbuh secara efisien di AI sambil juga membangun infrastruktur cloud AI untuk mendukung permintaan yang melonjak.
Harga minyak melonjak karena perang di Iran, berpotensi menaikkan biaya untuk membangun dan mengoperasikan pusat data. Dan di sisi produk, Copilot, asisten AI Microsoft, belum menunjukkan banyak traksi saat pengguna beralih ke layanan kompetitor dari Google, OpenAI, dan Anthropic.
"Redmond dalam kesulitan," tulis Ben Reitzes, analis di Melius Research, dalam catatan pada 23 Maret, merujuk pada markas Microsoft. Reitzes, yang memiliki rating hold untuk sahamnya, mengatakan perusahaan harus menggunakan kapasitas berharga dari cloud Azure-nya untuk memperbaiki Copilot, tetapi tidak memiliki pilihan "karena Copilot diperlukan untuk menjaga momentum di segmen paling menguntungkan dan terbesarnya."
Microsoft menolak berkomentar.
Sementara itu, saham perangkat lunak sedang dipukul-pukul sebagai bagian dari 'SaaSpocalypse' yang terinspirasi AI yang telah mendorong nama-nama seperti Adobe, Atlassian, dan ServiceNow turun lebih dari 30% tahun ini.
"Banyak dari SaaS tradisional sedang mati/mungkin dalam kemunduran terminal," tulis Jason Lemkin, pendiri SaaStr, minggu ini dalam postingan di X, menggunakan akronim untuk software as a service. Dalam postingan blog, dia mencatat bahwa multiple earnings untuk perangkat lunak tertinggal dari S&P 500.
Multiple Microsoft belum sekecil ini sejak kuartal empat 2022, ketika OpenAI memperkenalkan ChatGPT, menurut data Capital IQ.
Gil Luria, analis di DA Davidson, mengatakan ke CNBC bahwa jualan massal tidak dipertanggungjawabkan, dan dia merekomendasikan pembelian saham. Pada kuartal terakhir, Microsoft melaporkan pertumbuhan pendapatan hampir 17%, akselerasi dari setahun sebelumnya.
"Dislokasi antara kinerja fundamental Microsoft dan kinerja saham Microsoft, dan valuasi Microsoft, adalah terbesar dalam beberapa dekade," kata Luria. Dia mengatakan dia mengharapkan pertumbuhan laba perusahaan melampaui pasar yang lebih luas tahun ini.
"Tidak ada produk yang lebih lengket di semua perangkat lunak perusahaan daripada Microsoft Windows dan Office," katanya.
Microsoft telah mencoba membangun basis pendapatan yang lebih besar dari perangkat lunak produktivitas dengan add-on AI Microsoft 365 Copilot, tetapi sejauh ini hanya 3% pelanggan Office komersial yang memiliki lisensi untuknya. Luria mengatakan dia memiliki akses ke 365 Copilot, tetapi dia bukan penggemarnya. Yang lebih penting, katanya, Microsoft memiliki pricing power dengan langganan Office. Perusahaan mengumumkan rencana untuk menaikkan harga pada Desember.
'Demosi' Suleyman
Dengan Copilot berjuang untuk memenangkan pengguna, Microsoft mengatakan dua minggu lalu bahwa Mustafa Suleyman, mantan co-founder lab AI DeepMind yang sebelumnya memimpin pengembangan Copilot untuk konsumen, akan fokus membangun model AI. Microsoft telah menugaskan mantan eksekutif Snap Jacob Andreou untuk memimpin pengalaman Copilot untuk konsumen dan klien komersial.
"Ada kekhawatiran bahwa bisnis Microsoft 365 Copilot tidak memenuhi ekspektasi mereka, dan itu adalah area yang bisa melihat kompetitor baru," kata Kyle Levins, analis di Harding Loevner, yang memegang $219 juta saham Microsoft pada akhir Desember.
Levins mengambil perubahan struktur yang melibatkan Suleyman sebagai kabar baik. Yang lain tidak.
"Tentu terdengar seperti demosi pada yang terbaik," mantan trader Jane Street Agustin Lebron menulis di X. Perubahan ini mengikuti perginya eksekutif terkemuka, termasuk kepala gaming Phil Spencer dan Rajesh Jha, pemimpin produktivitas tertinggi Microsoft yang akan pensiun.
Microsoft masih mendapatkan pertumbuhan yang sehat dari Azure, yang berada di posisi kedua setelah Amazon Web Services dalam infrastruktur cloud. Pendapatan di divisi tersebut melompat 39% pada kuartal Desember. CFO Amy Hood mengatakan pada Januari bahwa pertumbuhan bisa berada di angka 40-an jika perusahaan mengalokasikan semua chip AI-nya ke Azure, daripada memberikan beberapa kepada tim yang mengoperasikan layanan seperti Microsoft 365 Copilot.
Azure mendapat manfaat dari backlog besar bisnis dari OpenAI dan Anthropic. commercial remaining performance obligations Microsoft di Azure lebih dari dua kali lipat pada kuartal Desember dari setahun sebelumnya menjadi $625 miliar.
Ini adalah pengingat bahwa, di antara hyperscaler tech, Microsoft dilihat sebagai pelopor awal dalam AI generatif karena investasi 2019-nya di OpenAI dan kemitraan strategis dengan startup tersebut. Namun perusahaan-perusahaan tersebut tidak lagi memiliki pengaturan eksklusif ketika itu comes to infrastruktur cloud, dan sekarang bersaing di sejumlah area.
Pada Februari, OpenAI mengumumkan layanan bernama Frontier yang perusahaan sebut "membantu perusahaan membangun, menyebarkan, dan mengelola AI agent yang bisa melakukan pekerjaan nyata."
CEO Microsoft Satya Nadella telah memakai wajah berani, mempromosikan peningkatan AI perusahaan di media sosial.
"Ini banyak persaingan ketat, tetapi tidak sepenuhnya zero-sum, seperti yang beberapa orang katakan," katanya pada Januari.
Aaron Foresman, managing director equity research di Crawford Investment Counsel, investor Microsoft, mengatakan kehadiran Nadella yang terus-menerus sangat penting untuk perusahaan yang dia pimpin sejak menggantikan Steve Ballmer pada 2014.
"Kami memiliki banyak kepercayaan dan keyakinan pada Satya," kata Foresman.
LIHAT: Tal Liani Bank of America bicara mengembalikan Microsoft sebagai 'beli'
Diskusi AI
Empat model AI terkemuka mendiskusikan artikel ini
"Microsoft's valuation compression is disconnected from its fundamentals—17% revenue growth with pricing power and a $625B Azure backlog doesn't warrant a 23% quarterly drop; the market is pricing in Copilot failure as existential when it's actually a rounding error against Office/Azure."
The article conflates two separate problems and overstates both. Yes, Copilot adoption is weak (3% of Office customers)—that's real. But Azure grew 39% with a $625B backlog, and Microsoft raised prices on Office despite Copilot struggles, suggesting pricing power remains intact. The 23% Q1 drop appears to be multiple compression, not fundamental deterioration: 17% revenue growth is solid. The 'SaaSpocalypse' framing is hyperbolic—traditional SaaS isn't dying, it's normalizing after 2021 excess. The article treats AI capex as a pure cost; it's actually a moat if Microsoft's infrastructure advantage holds. Suleyman's move reads as organizational optimization, not crisis.
If Copilot's 3% adoption signals that AI-enhanced productivity software has no real user demand (not just execution issues), then Microsoft's massive Azure capex spend becomes a stranded asset. And if OpenAI/Anthropic can now shop their workloads to AWS or Google Cloud, Microsoft's $625B backlog is less defensible than it appears.
"The market is mispricing Microsoft as a struggling software vendor rather than the foundational infrastructure provider for the enterprise AI transition."
The market is conflating a 'Copilot adoption' problem with a structural 'Microsoft' problem, creating a massive valuation disconnect. While the 3% penetration rate for Copilot is underwhelming, the article ignores that Azure’s RPO (Remaining Performance Obligations) hit $625 billion—a staggering indicator of long-term enterprise lock-in. Microsoft isn't just selling software; they are the infrastructure layer for the entire AI transition. The 'SaaSpocalypse' narrative is overblown; Microsoft’s pricing power in Office 365 remains unmatched. At current multiples, the market is pricing in terminal decay, yet the company is growing revenue at 17% with high-margin cloud dominance. This is a classic mispricing of a core utility.
If Azure's massive RPO is actually driven by pass-through costs for OpenAI and Anthropic rather than high-margin internal software, Microsoft's margins will compress significantly as they subsidize their competitors' compute needs.
"Microsoft’s stock weakness is driven by legitimate execution and margin risks around AI monetization, not by a collapse of its core business, so outcomes hinge on whether management can convert Azure scale and Office pricing power into profitable AI revenue before costs and competition bite."
Microsoft’s share-price shock reflects a concentrated investor worry: can the company turn AI hype into durable, profitable revenue before competition and rising infrastructure costs erode margins? Facts cut both ways — Azure grew ~39% last quarter and Microsoft reports $625B of commercial remaining performance obligations, yet Copilot adoption is only ~3% of commercial Office customers and the stock dropped 23% in the quarter. Key risks the article underplays: RPO timing/quality (multi‑year bookings aren’t immediate revenue), potential margin pressure from dedicating scarce AI chips to low‑paying products, and rising energy/capex if oil prices stay elevated. Leadership changes and fierce cloud/AI competition make execution the decisive variable.
This pullback may be an overreaction: Azure momentum, pricing power in Microsoft 365, and deep pockets to buy/scale models mean upside if execution normalizes — selling now risks missing a re‑rating once AI monetization shows up in results.
"Azure's doubled $625B backlog guarantees sustained AI cloud growth, dwarfing Copilot's nascent hurdles."
MSFT's 23% Q1 drawdown (vs. Nasdaq's 7%) reflects AI anxiety, but ignores accelerating 17% revenue growth, Azure's 39% surge (40s sans Copilot diversion), and $625B commercial backlog (doubled YoY)—a multi-year tailwind from OpenAI/Anthropic deals. Copilot's 3% adoption is early-stage; Office pricing power (hikes announced Dec) and Windows stickiness intact. Valuation trough since ChatGPT launch undervalues 15%+ EPS outlook. Oil at $85+/bbl adds datacenter capex pressure, but hyperscalers pass-through costs effectively.
Copilot's weak traction forces Azure capacity trade-offs, risking margin erosion if energy costs spike further from geopolitics and competition from Google/Anthropic accelerates.
"The $625B RPO is only bullish if Microsoft's own software margins, not commodity compute pass-throughs, drive it."
Everyone's anchoring on the $625B RPO as proof of durable demand, but Grok and Gemini gloss over a critical distinction: how much of that backlog is actually *Microsoft's* high-margin software vs. pass-through Azure compute costs for OpenAI/Anthropic? If it's 60%+ the latter, the RPO is a revenue mirage masking margin compression. ChatGPT flags this risk but doesn't quantify it. That's the real stress-test nobody's done.
"The $625B RPO is not just a margin risk, but a major antitrust liability that could force structural changes to Microsoft's integrated AI business model."
Claude is right to question the RPO quality, but everyone is missing the regulatory tail risk. If the DOJ or FTC labels Microsoft’s integration of OpenAI into Azure as a 'tying' arrangement, that $625B backlog becomes a massive antitrust liability rather than a moat. Investors are pricing for a tech cycle, but they should be pricing for a structural breakup risk. Microsoft's 'infrastructure advantage' is currently being viewed as a security blanket, but it's actually an antitrust bullseye.
"Undisclosed customer concentration in Azure RPO (large AI tenants) is a material downside risk that the panel hasn't sufficiently stressed."
We're fixated on antitrust and RPO quality but missing single-customer concentration risk: a handful of AI tenants (OpenAI, Anthropic) may drive a disproportionate share of Azure’s incremental demand. If top customers represent an outsized slice of near-term RPO, a renegotiation, migration, or pricing fight could quickly erase the margin story. Microsoft should disclose RPO concentration and pass-through vs. retained margin; absent that, downside tail risk is understated.
"Customer concentration in Azure is low at <15% for top 10, mitigating tail risks, though energy costs pose a clearer margin threat."
ChatGPT's concentration risk is valid but incomplete—Microsoft's top 10 customers are <15% of Azure revenue per filings, diversified across enterprises, not just AI labs. OpenAI/Anthropic are growth drivers (10-20% of incremental), but multi-year contracts and switching costs deter migration. The overlooked angle: if oil hits $100/bbl from MidEast tensions, datacenter power (20% of opex) spikes 15-20% without full pass-through, compressing margins 200-300bps.
Keputusan Panel
Tidak Ada KonsensusThe panelists debate Microsoft's valuation and growth prospects, with most acknowledging Copilot's slow adoption but arguing that Azure's growth, pricing power, and long-term backlog outweigh this concern. Key risks include the quality of Azure's remaining performance obligations, potential antitrust scrutiny, and concentration of top customers.
Azure's strong growth and the long-term backlog, which could provide a multi-year tailwind for Microsoft's revenue.
The quality and composition of Azure's remaining performance obligations, particularly the extent to which it represents high-margin Microsoft software versus pass-through Azure compute costs for OpenAI/Anthropic.